Study population and data collection

This was a single-center, case-control study. We enrolled patients with either newly implanted CIED (PM, ICD, CRT) or who underwent battery replacement at our hospital from May 2016 to January 2020. The inclusion criteria were as follows: patients over 65 years of age who underwent de novo CIED implantation or battery replacement and were treated with cardiac rehabilitation. Individuals < 65 years of age and those with missing data were excluded.

For all participants, we collected data on age, sex, diagnosis comorbidities, medical history, echocardiographic findings, laboratory data, operative records, and routine medical procedures from the electronic medical record system. Preoperative frailty was calculated using the mFI, based on medical history at the time of admission and current medical history as recorded in the electronic medical records. We divided the participants into groups according to their mFI results. Based on previous studies, those with two or fewer mFI items were placed into the non-frailty group, while those with three or more items were placed into the frailty group. The primary outcome was defined as follows: readmission for any cause deemed to require inpatient care within one year of treatment; the maximum follow-up period was regulated per the year.

Statistical analysis

We conducted unpaired t-tests for the continuous variables, which were expressed as means ± standard deviations. We employed Fisher’s exact test for the nominal scales, with results expressed as N (%) or medians (min-max). Next, we conducted the Cox proportional hazards regression with one-year readmission set as the objective variable. Based on the outcomes of previous studies, we selected the following dependent variables: sex [19, 20], age at implantation (divided into 75 years of age or older and less than 75 years of age [19, 20]), left ventricular ejection fraction (more than 50%; heart failure with preserved ejection fraction, more than 40%, less than 50%; heart failure with mid-range ejection fraction, and less than 40%; heart failure with reduced ejection fraction [HFrEF]) [21], kidney failure (estimated glomerular filtration rate < 60%) [22], anemia (hemoglobin < 10 g/dl) [23], hyponatremia (sodium < 135 mEq/L) [24], ischemic or non-ischemic [25–27], type of CIED [28], type of implant, and frailty [7]. In cases where variables were significant, we created Kaplan-Meier survival curves and conducted intergroup comparisons using Kaplan-Meier estimators with log-rank tests. P-values < 0.05 were considered statistically significant.

We conducted all statistical analyses using EasyR version 1.52 (Saitama Medical Center, Jichi Medical University, Saitama, Japan), which is a graphical user interface for R version 4.02 (The R Foundation for Statistical Computing, Vienna, Austria). More precisely, it is a modified version of the R commander which is designed to add statistical functions that are frequently used in biostatistics [29].

This study was approved by both the Ethics Committees of Saitama Medical University International Medical Center (20–215) and Niigata University of Health and Welfare (18692–210722), and was registered with University hospital Medical Information Network (identification number: 000044925). The requirement of informed consent from patients was waived, and we adopted an opt-out method, which allows patients to express their desire to not participate. It was approved by both ethics committees. The information regarding the use of medical record data for the study and the opt-out method is presented on our hospital’s website.

Major findings

First, the frailty group had a higher readmission rate than the non-frailty group. Indeed, our analysis showed that frailty was a significant predictor for readmission one year after surgery.

Second, the frailty group had significantly higher incidences of several mFI subitems when compared to the non-frailty group, including activities of daily living non-independence, diabetes, HF, myocardial infarction, and PCI/cardiac surgery/angina.

Third, there were many readmissions in the frailty group in each sub-analysis aimed at the presence/absence of HF. In particular, there was a significantly higher rate of readmission in non-HF patients with frailty.

Relationship between frailty and readmission in CIED patients

We investigated whether preoperative frailty affected postoperative readmission within a one-year period following implantation in patients with CIED. Our initial hypothesis was that CIED patients with preoperative frailty would have a higher postoperative readmission rate than those without frailty. In the study sample, 20% of CIED patients with preoperative frailty were readmitted for worsening HF within a one-year period following surgery. We also conducted univariate and multivariate regression analyses, the first of which included sex, age, cardiac function, renal function, anemia, hyponatremia, ischemia, type of CIED, and new implantation, and the second of which included cardiac function, type of CIED, and frailty as covariates. Patients’ sex has important research implications for frailty. For example, female frailty patients with HF have been shown to be at a higher risk of readmission [30]. However, in this study, the percentage of women was low as the subject of this study was heart disease. In addition, since mFI is an index that does not include sex, we consider sex to have had little effect on this study. The results of our multivariate analysis showed that frailty was significantly associated with readmission within a one-year period. In sum, these findings provide valuable insights for early efforts aimed at preventing readmission, which may be predicted prior to CIED surgery based on mFI results.

In comparison to the 20% of participants with frailty who were readmitted for worsening HF within one-year in this study, a previous study reported that 51.7% of participants who underwent CRT surgery were readmitted for HF within the same timeframe [6]. We believe this difference is due to variations in the participant characteristics and method of determining frailty. For example, the previous study was limited to patients with CRT, while this study considered a wider range of CIED, including PM, ICD, and CRT. In general, patients with PM and ICD have a lower rate of HF readmission and mortality when compared to CRT patients. The potential impact of physical and cognitive dysfunction on CRT is higher than those of PM and ICD. Frailty has been indicated as being an effect on readmission in CRT patients [13]. In particular, it has been reported that frailty was more frequent in non-responders to CRT [31]. Furthermore, there is literature suggesting a cognitive impairment in CRT patients [32]. Cognitive dysfunction has been an independent factor that increases readmission in heart failure [33]. In this study, only 16.8% of the current participants had CRT. When we added an investigation of the CRT non-responder rate in frailty and non-frailty group, we found that the rate of non-responders tended to be high in the frailty group, but there was no significant difference. Non-responders of CRT may be a cause of readmission for heart failure. CRT non-responders and frailty are covariates and difficult to separate because of a small number of objects. In addition, cognitive function was not investigated in detail in this study, but it may also attenuate the effects of CRT in the frailty population. However, we also found that frailty influenced one-year readmission rates after CIED surgery even after including cardiac function and the type of CIED as covariates.

Frailty, muscle, and CHF have a dynamic interrelation (Fig 4). Frailty and muscle are affected by activities of daily life, sensory impairment, cerebrovascular accident, peripheral vascular disease, and TIA. Some CIED patients with a background of cardiac disease are more prone to frailty. Hypertension, diabetes mellitus, myocardial infarction, angina, PCI/cardiac surgery, and chronic obstructive pulmonary disease cause frailty and are also direct causes of heart failure. Moreover, these induce organic disorder, which causes heart failure. Patients with frailty prior to CIED implantation have reduced activity, muscle strength, and power, which could further exacerbate heart failure after implantation. Reduced muscle mass or muscle function decreases muscle pump function and increases the risk of heart failure. Heart failure not only reduces cardiac function, but also contributes to reduced exercise tolerance, making the patient more susceptible to fatigue, which further contributes to heart failure symptoms.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Correlation chart for frailty, heart failure, CIED patients, and other mFI factors.

Frailty and muscle are affected by ADL, sensory impairment, CVD, PVD, and TIA. Frailty and CHF are affected by hypertension, diabetes mellitus, MI, anginal pain, PCI, and cardiac surgery. Muscle and frailty are associated with decreased muscle mass and strength and decreased activity. Muscle and CHF are related to exercise tolerance and muscle pump action. Frailty and CHF are related to cardiac dysfunction and fatigue. Abbreviations are the same as in Table 1 and Fig 1.

https://doi.org/10.1371/journal.pone.0277115.g004

Usefulness and bias of mFI

While this study used the mFI to determine frailty, previous studies have used the Fried method [34]. Of note, the mFI can be used to determine the presence or absence of frailty based on medical records; that is, without the need for special equipment. Since the Fried frailty assessment [34] requires a walking speed measurement, it is not considered suitable for preoperative screening among CIED patients who are at risk of syncope or worsening HF. On the other hand, the mFI is both practical and safe for predicting one-year readmission among this population. However, it is worth mentioning that the mFI subitems are biased toward information related to heart disease. Further, very few studies have investigated the effects of frailty on readmission rates within one year after CIED surgery, which highlights the need for continued investigation.

As three of the 11 mFI subitems are related to heart disease, including HF, practitioners and researchers should consider this bias when assessing frailty. For that reason, we examined the characteristics of participants both with and without HF histories. While we believed that frailty would influence readmission in both cases, this tendency was significant in non-HF participants. However, the same participants may have also been affected by renal function, anemia, the type of CIED, and usage of HF medication. Therefore, previous studies may have rather evaluated frailty. Patients with high mFI seemed to be prone to readmission in the hospital with or without a CIED. Future studies should increase the number of patients and consider the differences between patients with and without a CIED. In this study, the multivariate analysis was hindered by the small number of events, which also highlights the need for continued study.

Clinical implications

This study produced three important results for consideration in the clinical context. First, we found that the mFI could safely be used to predict one-year readmission prior to CIED implantation. Second, we believe that our results could help avoid futile treatment. Third, preoperatively applied mFI can be used to determine the risk of readmission early in the rehabilitation process. This information may be particularly applicable in revealing patients who should be targeted for cardiac rehabilitation, including exercise therapy and lifestyle guidance. Future studies should also search for ways to mitigate fragility in CIED patients, thus reducing the potential for HF readmission.

Study limitations

This study also has some limitations. For one, its retrospective approach was suitable for predicting one-year readmission rates based on preoperative frailty, but there was no consideration of postoperative status. There was also a limit pertaining to the statistical processing of the results due to the small number of events.

The most noteworthy problem was the differences in patient characteristics between the frailty and non-frailty groups and rates of sex, in addition to the small number of patients and events. The frailty groups had poorer cardiac function, higher rates of ICD and CRT implantation, and higher use of β blockers, mineralocorticoid receptor antagonist, diuretics, and amiodarone; further, this was biased toward the HF population. For that reason, we investigated the effects of frailty on the HF and non-HF groups. In the frailty group, both the HF and non-HF groups showed a tendency toward readmissions, but this was not significant in the HF group. Another difference in patient characteristics is sex. Female frailty patients with HF have been shown to be at a higher risk of readmission [30]. The patients in our study were predominantly male. Thus, we cannot rule out that sex may have influenced readmission. Hence, future studies should predict readmission using the mFI among larger samples, controlling for patient characteristics. Furthermore, it may be necessary to develop an advanced index that excludes the association with heart disease.

In addition, the frailty group had higher rates of diabetes, previous myocardial infarction, and previous PCI/cardiac surgery/previous angina. In this regard, past research has shown that factors such as diabetes mellitus, myocardial infarction, and history of PCI may increase the risk of HF readmission in HF patients [21], and may thus be confounding factors. However, these studies did not evaluate frailty in this context.

Finally, the mFI is calculated based on items related to the patient’s medical history, meaning there is little room for improvement after surgery. This indicates the need for postoperative evaluations aimed at determining whether preoperative or postoperative frailty and cognitive dysfunction impact readmission.